The Q4 2017 growth of the server market continues in Q1 2018, Gartner reports-- EMEA revenues are up by 32.1% Y-o-Y even as shipments drop  by -2.7 Y-o-Y, with hyperscale, enterprise and midsize datacentres as main spending drivers.

"Enterprises and midsize businesses are in the process of investing in their on-premises and colocation infrastructure to support server replacements and growth requirements even as they continue to invest in public cloud solutions," the analyst says. "Additionally, when it came to server average selling prices (ASP) increases for the quarter, one driver was the fact that DRAM prices increased due to constrained supplies."

Hitachi subsidiary Hitachi Vantara adds NVMe SSD caching and GPU-based server options to the Unified Compute Platform (UCP) Line of converged, hyperconverged and rack-scale systems.

The company also has certified application ecosystem solutions optimised for mission-critical environments, including SAP HANA, Oracle databases, VMware and big data analytics frameworks. UCP system are based on pre-tested modular blocks of infrastructure validated to meet specific needs, allowing customers to start small and grow systems with increments of compute, network and storage components.

Nvidia presents an upgrade on the HGX-1 hyperscale GPU accelerator chassis-- the HGX-2, a server powered by no less than x16 Tesla V100 GPUs and x12 NVSwitches, making "the world's largest GPU."

According to the company, the HGX-2 is “the first unified computing platform for both artificial intelligence and high performance computing.” It offers up to 2 petaFLOPS performance for low-precision tensor operations, 250 teraFLOPS at single-precision and 125 teraFLOPS for double-precision.

The NVSwitches allow GPU-to-GPU communications reaching up to 300GB per second, and topping everything of is 0.5TB of GPU memory. The result is a unified system developers can consider as a single, huge GPU. It can handle a variety of tasks, including cloud, AI, deep learning or high-performance simulations, and Nvidia claims a single HGX-2 unit can replace 60 CPU-only servers.

Global server revenues are up by 38.6% Y-o-Y in Q1 2018, IDC reports, reaching $18.8 billion and marking a 3rd consecutive quarter of double-digit growth. Shipments total 2.7 million, a 20.7% Y-o-Y increase.

EMEA revenues show 35% Y-o-Y growth.

Such "historic" server demand is the result of a market-wide enterprise refresh cycle, strong demand from cloud service providers, increased server use as "core building blocks" of software-defined infrastructure, broad demand for new CPUs (such as the Intel Purely platform) and growing next generation workload deployments. ASPs are also on the up during Q1 2018 thanks to richer configurations and increased component costs, further driving revenue growth.